ambient composition sessionheptane as a diesel surrogate. there are however noticeable differences...

ECN1 Workshop 2011 Ambient Composition Section – Jaclyn Nesbitt of 2

Ambient Composition Session

Group Leader: Jaclyn Nesbitt (Michigan Technological University - MTU)

Contributors: Maarten Meijer (Eindhoven University of Technology), Louis-Marie Malbec and Gilles

Bruneaux (IFPEN), Lyle Pickett (Sandia National Laboratory), Raul Payri (CMT), Tim Bazyn (Caterpillar)

Summary of the Ambient Composition Working Group Session

The focus of the ambient composition working group for the ECN1 Workshop was to examine

the different methods of creating the thermodynamic state for the charge gas in experimental

apparatuses including combustion vessels and constant pressure flow rigs in the Engine Combustion

Network (ECN) under the ‘Spray A’ test condition. This was accomplished by comparing the different

institutions ambient compositions and modeling their impact on autoignition of n-heptane as a diesel

fuel surrogate. The slides presented at the workshop are provided, and a brief overview of the

comparisons made with the key results outlined here.

First, the four preburn combustion vessels (Sandia, MTU, IFPEN, and Eindhoven) were compared

in regards to preburn mixtures used which consist of varying levels of C2H2, C2H4, H2, O2, N2 and Ar. This

comparison was accomplished using a single-zone chemical kinetics model in Cantera interfaced with

Matlab with the GRI 3.0 mechanism to compare minor species produced during the preburn and levels

at injection. The different cool-down temperature histories of the vessels (heat transfer modeled as

convective with a temperature trace being a quadratic exponential decay) were also modeled. The peak

preburn temperature of Eindhoven is largest, followed by MTU, Sandia, and IFP. Although MTU and

Sandia utilize the same preburn mixture, MTU uses a fan speed seven times that of Sandia which results

in a higher peak temperature and a faster rate of cool down. MTU and IFP have similar cool-down rates,

with Sandia having the longest cool-down time due to the low fan speed used. These differences

(premixed gas mixture and cool-down rate) cause variations in minor species, NO is largest for

Eindhoven, followed by MTU, Sandia and IFP due to the temperature trends from the preburn, but in all

cases the levels at the time of diesel injection are more than 50 times less than equilibrium levels, with

similar trends for NO2. For OH, this species tracks the temperature-time trace and at injection IFP has

the maximum OH (attributed to the largest level of H2 in the preburn mixture), followed by Eindhoven,

MTU and Sandia, with these levels being one to two orders of magnitude less than equilibrium values. At

equilibrium, for IFP NO in 3030 ppm and OH is 170 ppm, for MTU and Sandia NO is 5100 ppm and OH is

230 ppm, and for Eindhoven NO is 5250 ppm and OH is 240 ppm.

Additionally, for the conditions at injection the four preburn vessels major species (CO2, H2O, Ar,

N2, and O2) were compared to the CMT and Caterpillar constant pressure flow rigs which use O2 and N2

to achieve the desired 15 percent oxygen conditions for ‘Spray A’. Comparison included the major

species levels and constant pressure specific heat capacities. Finally, the stoichiometric n-heptane

ignition delay was determined using chemical kinetics modeling with a reduced mechanism to compare

the influence of ambient compositions major species levels and the different minor species produced

during the preburn on autoignition, relative to that of dry air and air plus ideal residuals to simulate EGR.

Results showed that the preburn vessels, despite their respective differences in minor species at

injection, did not significantly impact the ignition delay of n-heptane. More specifically, the vessel

ignition delays were 4% shorter than ideal EGR and 83% longer than dry air. Comparing the preburn and

constant pressure vessels major species at fuel injection, the specific heat capacities are similar

(spanning 1.13 to 1.21 kJ/kg-K at 900 K), and the variation in ignition delay is within the modeling

accuracy. More specifically, when considering major species only, IFP had the shortest ignition delay

(0.65 ms), followed by CMT / Caterpillar at 0.69 ms, and the remaining preburn vessels of 0.72 ms.

Again, the ignition delay was longer than that of dry air as expected due to the reduction in oxygen level

at the ‘Spray A’ condition. There are however noticeable differences in the peak temperature of n-

ECN1 Workshop 2011 Ambient Composition Section – Jaclyn Nesbitt of 2

heptane ignition between the different vessels with the Nitrogen / Oxygen constant pressure flow rigs

having a peak temperature at 45 K higher than the preburn vessels. Key conclusions are that despite the

differences in preburn mixtures and ambient compositions, the ambient charge-gas compositions for

the vessels considered show no significant differences in fuel ignition delay at the condition considered

or mixture constant pressure specific heat capacity.

Also discussed briefly was a recent Energy and Fuels journal publication with a goal of isolating

and understanding chemical kinetics effects on the preburn process and autoignition of n-heptane fuel

while including minor species and neglecting spray dynamics. The key conclusions can be found in the

slides, with the main result being that the combustion vessel with the preburn procedure is an effective

tool for studying spray combustion over a range of ambient conditions without concern over the

reactive minor species produced by the preburn procedure. The citation is listed below for reference for

further review.

Nesbitt, J.E., Johnson, S.E., Pickett, L.M., Siebers, D.L., Lee, S-Y., Naber, J.D., ‘Minor Species

Production from Lean Premixed Combustion and Their Impact on Autoignition of Diesel

Surrogates,’ Energy and Fuels 25 (3), pp. 926-936, 2011.

Discussion Issues & Questions

The key observation and conclusion from the above and the workshop is that there is no

significant effect of ambient composition differences on the specific heat and autoignition delay of n-

heptane as a diesel surrogate. There are however noticeable differences in the flame temperatures and

the consequences of this are unknown. Further unknown is the potential differences in ignition or

combustion duration as a result of these ambient environments. Questions were brought up on if the

time-step used in the model influences the results (a consistent time-step was used in all of the

modeling, however, there was no sensitivity study undertaken), the influence of other n-heptane

stoichiometries in addition to the lambda one case studied here, and if the n-heptane mechanism used

influences the results. These questions are continuously being addressed with additional modeling work.

Recommendations & Future Investigation

It is recommended that experimental testing is undertaken to characterize the differences in

ambient gas composition between the vessels in regards to major and minor species levels and

autoignition delay of the fuel. MTU has conducted some exhaust gas sampling, however this sampling

was not in-situ, not in real-time in the vessel, nor under ‘Spray-A’ conditions. Sandia has tried some

measurements of the exhaust gases but found that there are large levels of unburnt hydrocarbons due

to the large amount of crevice volumes in the CV (think of the CV as a cube with 6 pistons (access ports)

as the crevice volumes). Sampling directly in the CV is difficult but is something that would be useful

experimentally and to also validate this simplified chemical kinetics preburn modeling.

The modeling used in this work involved a simplified single-zone model, and using a more

detailed two-cell or multi-cell model would improve the accuracy and applicability of the modeling

simulations. LLNL expressed some interest in looking into this, but this will be extremely complex to the

differing vessel geometries.

1

ECN1 Workshop – May 13-14, 2011

Working Group Section:

Ambient Composition

Leader: Jaclyn Nesbitt, Michigan Tech University

Dr. Jeffrey Naber, Dr. Seong-Young Lee

Contributing Institutions: Sandia, Eindhoven, CMT,

Caterpillar, IFP, MTU

ECN1 – Workshop on Spray Combustion, Ventura, CA

May 13-14, 2011

1


Outline

• Details of MTU Combustion Vessel

• Chemical kinetics reactor modeling of ‘Spray A’ 15% oxygen

conditions for different institutions

� Comparison of major species at injection for different vessels

� Comparison of preburn environments

• Cool-down & minor species generation

� n-Heptane Ignition Delay

• Influence of major species

• Influence of prebrun

• Overview of Energy and Fuels Publication

� Nesbitt, J.E., Johnson, S.E., Pickett, L.M., Siebers, D.L., Lee, S-Y., Naber, J.D.,

‘Minor Species Production from Lean Premixed Combustion and Their Impact

on Autoignition of Diesel Surrogates’, Energy Fuels 25 (3) pp 926-936, 2011.

2

2


MTU Combustion Vessel Details

3

Laboratory Overview• Injection Systems

� Diesel (4140 Bar) / GDI (200 Bar)

� Piezo & Solenoid Injector Drivers

� Cooled injector windows

• Ignition Systems� Generic automotive coil

� Advanced ignition- dual-fan / spark plug

Remote Control and Monitoring System

Test Cell

3


Optical and Laser DiagnosticsDiagnostics

• Shadowgraph/Schlieren Imaging: Gas Phase

• Mie: Liquid phase (spray - droplet)

• TSI Phase Doppler Particle Analyzer (PDPA)–

Droplet sizing

• Rayleigh: Gas phase, air-fuel mixture

formation (ER), temperature

• PLIF: HCHO (cool flame)

• LII: Soot & PLIF: PAH

• Emissions: CH* & OH*

Cameras

• PCO DiCam Pro Intensified Camera (ICCD),

• PCO SensiCam Unintensified Camera (CCD),

• LaVision UltraSpeedStar16: 1 million

frame/sec

Laser and Light Sources

• 5 W Argon-Ion Laser

• High power Nd:Yag laser

• Short-pulsed high-power flashlamp


Diesel Spray Imaging

6

Liquid Phase – Back Scattering Imaging

High Pressure Common Rail Injector – 8 Holes; 2000 Bar Injection Pressure

34.8 kg/m3 Ambient Density

NonVaporizing Spray –

373 K N2

1.0 ms Injection Duration

Vaporizing Spray –

1200 K 0% O2


Combusting Spray –

1200 K 21% O2


Zoomed-In View

Vaporizing Spray - 1200 K 0% O2


4


Comparing ‘Spray A’ Ambient

Compositions

7


Goals & Objectives

• Goal:

� Examine test facilities charge gas composition impact on

autoignition in comparison to those in an engine through

kinetics modeling.

• Objectives:

� Compare major species at injection

� Compare preburn environments – cool-downs and minor

species produced

• Sandia / MTU / IFP / Eindhoven

� Compare n-heptane ignition delay under ‘Spray A’ conditions

1. Consider major species at fuel injection

2. Preburn minor species effects

8

5


MAJOR SPECIES COMPARISON AT

INJECTION

9


‘Spray A’ Ambient Environment Comparison – Post Preburn

‘Spray A’ Environment (Post preburn) –

Volume %

Institution

Ambient

Composition

Method

O2

(%)

N2

(%)

CO2

(%)

H2O

(%)

Argon

(%)

Mixture Specific

Heat (at 900 K)

(kJ/kg-K)

Sandia Preburn, Premixed 15.0 75.1 6.2 3.6 -- 1.16

MTU Preburn, Premixed 15.0 75.1 6.2 3.6 --- 1.16

IFP Preburn, Seq. Fill 15.0 71.7 1.7 11.6 --- 1.21

Eindhoven Preburn, Seq. Fill 15.0 71.2 6.4 3.2 4.2 1.13

CMT Flow Rig 15.0 85.0 -- -- -- 1.13

Caterpillar Flow Rig 15.0 85.0 -- -- -- 1.13

Ideal EGR – 38.3%

(CO2, H2O)-- 15.0 77.7 3.8 3.5 -- 1.16

Dry Air -- 21.0 78.1 -- -- 0.9 1.12

Modified Preburn

HCR (Match Diesel

R = 1.85)*

-- 15.0 75.3 6.1 3.6 -- 1.16

10

*Johnson, S.; Nesbitt, J.; Lee, S.-Y.; Naber, J. D. Journal of KONES, Powertrain and Transport 2009, 16 (2), P2-00-2.

6


COMPARISON OF PREBURN

ENVIRONMENTS

11


‘Spray A’ Ambient Composition Environment

Comparison – Preburn Mixture

Preburn Mixture Comparison (Volume %)

Institution

Ambient

Composition

Method

C2H2

(%)

C2H4

(%)

H2

(%)

O2

(%)

N2

(%)

Argon

(%)

Tadiabatic

(U,V)

(K)

SandiaPreburn,

Premixed3.06 -- 0.50 22.63 73.82 -- 1927

MTUPreburn,

Premixed3.06 -- 0.50 22.63 73.82 -- 1933

IFPPreburn,

Sequential Fill-- 0.816 9.39 21.43 68.36 -- 1772

EindhovenPreburn,

Sequential Fill3.15 -- -- 22.64 70.07 4.14 1946

12

Adiabatic flame temperature calculated for initial temperature and

pressure to match core density of ‘Spray A’ of 22.8 kg/m3

7


Comparison Experimental Preburn Data

13

Time 0 s = Spark Time


Chemical Kinetics Preburn Modeling Diagram

14

Constant-Volume Reactor

Output

• Mole factions of 53 Species

• TTarget

• PTarget

Premixed Combustion & Cool-Down

Software: Engineering Equation Solver (EES) & Cantera integrated into

Mathworks Matlab

Mechanism: Detailed GRI-Mech 3.0, 53 species, 325 elementary reactions

Extent of Reaction Calculation

Equate internal energies to elevate temperature

(T, P, Mole fractions of C2H2, H2, N2, O2, CO2, H2O, Ar)

Initial Conditions

(T, P, Mole

fractions of C2H2,

C2H4, H2, N2, O2,

Ar)

Initial Conditions Definition

8

15

Institution Time Constant Cool

Down Decay (s)

Fan Speed (RPM) Location

Sandia 6.5 1000 Upper Corner (opposite injector)

MTU 1.3 7000 Top Window

IFP 1.2 3140 Upper Corner (near injector)

Eindhoven 3.5 1890 Lower Corner

Cool-Down Curve Fitting

Time Shifted to 0 s at 900 K

Heat Transfer Modeling

– Curve fit exponential

decay function to

experimental data

(constants a, b, c):

– Assume convective heat

transfer, constant varied

to match exp. data

)exp()(2

cbtattT ++=

dt

)t(dTttanConsQ =

16

Preburn Modeling Results: NO and OH Time Traces

• At injection, Eindhoven maximum

NO (Highest peak temp preburn),

IFP minimum NO

• NO below equilibrium (IFP:

3000ppm, Sandia/MTU/Eindhoven

> 5000ppm)

• At injection, IFP maximum OH,

Sandia minimum OH

• OH levels below equilibrium (180

ppm IFP, 240 ppm Sandia / MTU /

Eindhoven)

NO – Time Trace OH – Time Trace

9


Preburn Modeling Results– Conditions at 900 K for

Fuel Injection

17

Sandia MTU IFP Eindhoven

Time (S) Relative to

Peak Cool-Down Temp2.15 0.95 0.55 1.68

Pressure (MPa) 5.81 5.82 6.13 5.71

NO (PPM) 11.24 59.14 0.11 96.72

NO2 (PPM) 2.47 9.64 0.02 16.66

OH (PPB) 3.78 5.24 14.34 3.87


N-HEPTANE IGNITION DELAY

COMPARISON

18

10


Fuel Autoignition Modeling – n-Heptane

19

Input

• Species Mole Fractions

• 53 GRI Species

(Preburn)

• Major Species (H2O,

CO2, N2, O2, Ar)

• Mix stoichiometrically

with n-heptane fuel

• Ttarget = 900 K (Spray A)

• Ptarget = ~6 MPa (Spray A)

Constant Pressure

Perfectly Stirred

Reactor, Constant

pressure / enthalpy

conditions

Output

• N-Heptane

ignition

delay

Ignition Modeling

Mechanism: Reduced, 179 species,

823 reversible reactions, modified

from LLNL to include NO and NO2

(from GRI-Mech 3.0)


Ignition Delay – Major Species Ambient Composition Comparison - Stoichiometric n-Heptane Mixture

20

Insignificant differences (within

modeling accuracy) in ignition

delay for different ambient

environments.

11


Autoignition Modeling Results – Influence of Preburn

21

Sandia MTU IFP Eindhoven

Air Plus Ideal

Residuals

(38.3%) –

15% O2

Dry Air

n-Heptane

Ignition Delay

(ms)

0.69 0.66 0.68 0.67 0.71 0.37

Peak

Temperature

(K)

2302 2302 2307 2320 2326 2707

• n-Heptane ignition delay at spray A conditions (900 K)

• No variation in ignition delay as result of preburn

• CV ignition delay 4% shorter than ideal EGR (15% O2)

• CV ignition delay 83% longer than dry air (21% O2)


Conclusions

• Major Species Comparison

� Similar specific heats

• Preburn Comparison

� Peak Temperature, NO and NO2 at Injection: Eindhoven , MTU

> Sandia > IFP

� OH at Injection: IFP > MTU > Eindhoven > Sandia

• Ignition Delay Comparison

� Major Species Consideration: no significant variation in n-

heptane ignition delay.

� Preburn Consideration: no variation in ignition delay due to

minor species from preburn

• CV Ignition delay 4% shorter than Ideal EGR (15% O2)

• CV Ignition Delay 83% longer than dry air

22

12


Minor Species Production from Lean Premixed

Combustion and Their Impact on Autoignition of

Diesel Surrogates

Energy and Fuels Volume 25, Issue 3, pp. 926-936

Jaclyn E. Nesbitt1, Samuel E. Johnson1, Lyle M. Pickett2,

Dennis L. Siebers2, Seong-Young Lee1, Jeffrey D. Naber*1

1Department of Mechanical Engineering – Engineering Mechanics

Michigan Technological University

Houghton, MI USA2Combustion Research Facility

Sandia National Laboratory

Livermore, CA, USA

23

ECN1 – Workshop on Spray Combustion, Ventura, CA, May 13-14 2011


Brief Overview: Goals and Main Conclusions• Goal: Isolate and understand chemical kinetics effect on preburn process

and autoignition of n-heptane fuel while including minor species, and

neglecting spray dynamics

• Conclusions:

� Preburn -- Most significant minor species formed is NO, OH tracks equilibrium, range of

NO produced in CV is not outside that representative of CI engines.

� Minor species of NOx and OH from preburn shorten ignition delay by 3% relative to air,

and increase it by 6% relative to air plus residuals at 1000 K. Impact is comparatively

insignificant due to accuracy of modeling simulations

� NO largest impact on igniPon: igniPon delay ↓ 10% for 10X increase in NO (120 ppm)

� CV useful for simulating heavy EGR-- CV ignition delay reduced by maximum 7% for

given percent O2 relative to ideal EGR with no minor species

� Change in minor species mole fraction – minimal influence on ignition delay, rather,

dominating trend is increase in oxygen yielding reduction in ignition delay

• Combustion vessel with preburn procedure is effective to study spray

combustion over various ambient conditions without concern over

reactive minor species produced by preburn.

24

13


QUESTIONS?

DISCUSSIONS


Summary, Recommendations & Further Investigation Required

• Summary / Recommendations

� No significant effect of ambient composition differences on specific

heat and ignition delay.

� Although preburn mixtures and cool-down histories are different

which yield differing minor species, this does not yield any significant

impact on autoignition of n-heptane as a diesel surrogate.

• Further Investigation Required / Guidelines

� Consequences of differences in flame temperature due to different

compositions.

� Experimental measurement of exhaust gas and species to validate

modeling and further compare the vessels

• If differences are significant relative to what was predicted by the model,

further assessment of the mixtures used will need to be undertaken to

standardize the preburn mixture to ensure consistencies for spray and

combustion studies.

26

ambient composition sessionheptane as a diesel surrogate. there are however noticeable differences...

Documents